Master of Engineering Science
Electrical and Computer Engineering
Dr. Ken McIsaac
A major component of the infrastructure of any modern city is a network of underground pipes that transport drinking water, storm water and sewage. Most of the pipes currently being used are made out of concrete or various plastics. As with any material, they have an expected lifespan after which deterioration begins to occur. This can result in cracks, and in some cases, even large holes in the pipe which can cause a complete loss of function of the pipe. These defects invariably lead to water losses that necessitate the repair of the pipeline, which is an expensive undertaking.
The purpose of this thesis is to give a detailed report of the development and testing of a robot with a spray head that is autonomously controlled. This spray head will deposit a liquid material onto the pipe that will then cure to form the new interior wall of the pipe. The design of the robot most suited to this task is a Stewart platform: a parallel manipulator that uses prismatic actuators to control a single end-effector. In contrast to the traditional Stewart platform design, which has six independently controlled legs that are used to control the position of the top platform, a novel design is used which has only three independently controlled legs. The advantages of this design are less weight, less complicated kinematics and a smaller design envelope. A circular trajectory was implemented in the microcontroller code and the accuracy of the Stewart platform was evaluated using videos and image processing techniques.
An optimization algorithm is proposed which combines the controlled random search algorithm and the particle swarm optimization algorithm. The effectiveness of this algorithm is demonstrated by selecting the design parameters of a 3-DOF Stewart platform so that the radius of the circular spray path is maximized.
Brecht, Derek K., "A 3-DOF Stewart Platform for Trenchless Pipeline Rehabilitation" (2015). Electronic Thesis and Dissertation Repository. 3156.